Passed 18 of 23 test cases with the skill, and 4 of 23 without it.
Without skill16%
With skill80%
Checks it passed that the model alone missed
Maps -Auto parameter to -Proceed
Directs the user to run /speckit.plan
Directs the user to run /speckit.specify
and 12 more in the full report
16 fixed4 passed either way5 still failing0 broke
Pass-rate lift
+64 pp
Tokens
+79%
Agent turns
0%
Response time
-57%
Model gemini-3.6-flashMeasured 8/11/2026v120 of 25 cases passed with skill · 23 comparable
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 16 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 16
Prompt
Where does the SpecKit updater save template backups prior to modifying files, and how many recent backups are retained?
What was checked, without and with the skill
Checked
Without skill
With skill
Identifies .specify/backups/ as the backup location
Missed
Passed
States that 5 recent backups are retained
Passed
Passed
case-16 · gemini-3.6-flash · 8/11/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md173 linessingle file
SpecKit Safe Update
This skill provides safe update capabilities for GitHub SpecKit installations, preserving customizations while applying template updates.
Installation: Before adding any marketplace or manual clone, inspect the
exact plugin revision and every bundled file, report scripts and network or
filesystem behavior, and ask for explicit user approval. Do not install a moving
branch directly into an active agent directory.
When to Use
You need to update or install SpecKit templates while preserving project customizations.
You want a safe approval flow around update, rollback, or version-specific SpecKit operations.
The task is to operate the SpecKit updater conversationally instead of running raw commands blindly.
What to do when this skill is invoked
When the user invokes /speckit-updater, you should:
Run the update orchestrator script without any flags (conversational mode):
Audit SQL for the cost & performance anti-patterns that burn warehouse credits. Scores warehouse health 0-100 and outputs a prioritized cost-reduction plan for BigQuery, Snowflake, Redshift, and Postgres.